UNIVERSITY OF NORTH CAROLINA AT CHARLOTTE

Department of Electrical and Computer Engineering

Experiment No. 8 -Switch Debouncing

Overview

This experiment will highlight one of the problems encountered in digital systems design, that of
switch bouncing. In this laboratory an example is provided to illustrate the full effect of this
problem and its solution, a switch-debouncing circuit. The circuit, as shown in Figure 4, makes
use of a cross-coupled NAND gate SR latch. By inserting this latch into the clock-push-button
circuit, bounce pulses, generally of a relatively high rate, will not move through the SR latch.

An important (but often overlooked) concern in digital design is switch bounce. The basic
mechanisms involved are related to the mechanical design of the switch and the large electric
fields that develop between the contacts when they are very close but not touching. The result of
this situation is arcing between the contacts until they finally settle down and make permanent
contact. In many design situations, this arcing (or bouncing as it is more often referred to) is of
no concern to the system operation. In other situations, however, this switch bouncing can cause
seriously undesirable results.

As an example where we see the full effect of the switch bounce problem, consider the circuit
illustrated in Figure 1. In this circuit, we use a simple push button to provide the clock input
signal to a binary counter and another push button switch to provide the clear input signal. The
counter is a 74161, 4-bit, binary “up” counter, which is leading edge-triggered with an active low
asynchronous clear and an active high synchronous parallel load. To disable the parallel load this
input is connected to ground. Observe that we use a 10kΩ resistor to “pull-up” the clock, and
clear inputs to the counter. With these switches is in the open position a logical 1 is applied and
in the closed position a logical 0 is applied. It is good design practice to always pull-up (to a
logic 1) or pull-down (to a logic 0) critical unused or unconnected inputs. The unwise alternative
would be to let such inputs “float” and expect that they will always assume and remain at the
desired level.

Figure 1. Example Circuit

When the Clear button is pushed, the clear input which is active low is pulled to ground (logical
zero), and the counter resets to the all zeros state. Since the clear input to the counter is
asynchronous, clearing happens automatically without a clock pulse. This is the appropriate
behavior by the system and we see no effects of switch bouncing (repeated low going pulses to
this input would do nothing more than clear the counter again and again). The desired behavior
of the counter with respect to the Clock pulse is that no change in the output (Q0 through Q3) is
observed on the trailing edge when the button is pressed in, but each time the button is released,
causing a leading edge, the counter will increment by 1. As illustrated in Figure 2, if switch
bounce or arcing is present, the desired operation is prevented and multiple leading edges will
cause the counter to increment by an unknown amount as the clock input bounces back and forth
between low and high. When final contact is made the clock remains at a logical zero until the
switch is opened; but then, arcing begins and continues until the contacts are far enough apart to
prevent arcing. Thus, we see multiple rising edges at both the closing and opening of Clock
button, and as a result, an unknown accumulation of counts by the counter.

Figure 2. Clock signal and output response

The solution to this problem is a switch de-bouncing circuit, which makes use of an SR latch
constructed from cross-coupled NAND gates as shown in Figure 3, along with its associated
state table. By inserting the SR latch into the Clock push button circuit, as shown in Figure 4, we
can overcome the bouncing problem by making use of the Set, Reset, and Storage states of the
SR latch. Notice that we have also changed the type of switch from a single-pole, single-throw
switch to a double-pole, single throw switch, and thus, have needed to add an additional pull-up
resistor. The double-pole switch ensure that the S and R inputs to the SR latch will both be at a
logical 1 (putting the SR latch in the storage state) when no contact is made with either pole.
This is the case while we are in the process of pushing or releasing the switch.

2
 Figure 3. SR latch schematic and state table

Comparing the action of the switch to the state table for the SR latch we can see how the switch
debouncing circuit works to prevent glitches in our clock signal. Assume that the switch is
normally connected to the S (set) input of the SR latch as illustrated in Figure 4. As a result, the
S input is pulled to a logical 0 and the Q output to a logical 1, which is the Set state of the SR
latch. When the clock button is pushed, the break in contact takes the SR latch inputs from S=0
and R=l to S=1 and R=l, corresponding to the storage state where the Q output continues to be a
logical 1. Glitches at the S input of the SR latch will take the latch between the set state and the
storage state, keeping the Q output at a logical 1. As the switch arm moves toward the contact
connected to the R input, the latch remains in the storage state with Q = 1 until it encounters the
first arc between the switch arm and the R contact. With this first arc, the inputs to the latch
change from S=1 and R=1 to S=1 and R=O, corresponding to the reset state which produces a
logical 0 at the Q output. Subsequent glitches take the RS latch between the storage state and the
reset state keeping Q unchanged. As a result, we see a clean falling edge of the clock input to the
counter as illustrated by the Expected Clock in Figure 2.

Figure 4. Switch debouncing circuit

When the button is released, the SR latch moves between the reset and storage states as arcing
glitches continue until sufficient distance is achieved and the arcing stops. As the switch arm
nears the contact connected to the S input, arcing begins, and the latch moves between the set
and storage states with the first glitch producing a logical 1 at the Q output. This transition in Q
produces a rising edge of the clock signal and the counter is incremented once and only once
until the next press and release of the clock push button.

Note that the double-pole switch at the inputs of the SR latch prevents the latch from entering the
indeterminate state where both S and R are 0 and both the Q and Q bar outputs are forced to a
logical 1. In this situation the final state of the latch is indeterminate and depends on what is
referred to as a race condition. Furthermore, a conflict of this type could, in some cases, result in
circuit damage.

3
 Pre Lab – Switch Debouncing

Given the logic diagrams in Figures 1 and 4, along with the appropriate data sheets for the 74161
(binary counter) and 7400 (quad 2-input NAND gates) integrated circuits, draw a schematic for
each of these two circuits. Most importantly, be sure to label the IC pin numbers since you will
be using these schematics during the lab session. Show your pre-lab schematics to the lab TA at
the beginning of the lab session. Note that the control inputs LOAD and CLR are active low.
Thus, they must be connected at a logical high (5V) to be inactive.

(INSTRUCTOR’S SIGNATURE________________________________DATE___________)

4
 Lab Session — Switch Debouncing

1. Show your pre-lab schematic to the lab TA at the beginning of the lab session.

2. Construct the circuit without debounced switch (Figure 1) on the prototype board.
Connect the counter outputs to LEDs on the prototype board in order to monitor state of
the counter during the exercise.

3. Clear the counter by pressing the Clear push button and verify that the counter goes to the
all 0s state.

4. Press the Clock push button but do not release the push button and record the count value
with the push button depressed along with the value you would expect for the counter if
there were no switch bouncing problems.

5. Release the Clock push button and record the count value after the push button is released
along with the value you would expect for the counter if there were no switch bouncing
problems.

6. Clear the counter and verify that it has reset to the all 0’s state. Then press and release the
Clock push button and record the count value. Repeat this process of clearing the counter,
then pressing and releasing the Clock push button a number of times (at least 5 times)
while recording the count value each time. Try pressing and releasing the button slowly
as well as fast to see if there is any difference in the count values. Take the average of the
count value obtained for the press/releases of the Clock push button.

7. Construct the switch debouncing circuit and integrate it with your counter to obtain the
circuit in Figure 2. Repeat Steps 3 through 6 for the debounced clock circuit recording
values as indicated in each step.

8. Before you disassemble your circuit and leave the lab, show your lab TA that your
debouncing circuit works properly.

5
 Lab Session – Measurements Lab (Data Sheet)

Table 1: With no debouncing switch

Trial Expected count value Measured count value

# LED 1 LED 2 LED 3 LED 4 LED 1 LED 2 LED 3 LED 4
1
2
3
4
5
6
AVG

Table 2: With debouncing switch

Trial Expected count value Measured count value

# LED 1 LED 2 LED 3 LED 4 LED 1 LED 2 LED 3 LED 4
1
2
3
4
5
6
AVG

TA viewed debouncing circuit and works properly:

(TA initialize)

INSTRUCTOR'
S INITIALS DATE:

6
 Post Lab — Switch Debouncing

1. Correct any mistakes in your two schematics that you encountered during the
construction and operation of your two circuits (with and without the switch debouncing
circuit). Prepare the data you took during the lab session (including the count values and
averages) in tabular form and write a brief discussion of the effectiveness of the switch
debouncing circuit based on the data.

2. Given that the debouncing circuit is somewhat expensive in terms of hardware (2 NAND
gates, 2 resistors, and a double-pole, single throw switch), describe applications where
you would require switch-debouncing circuits as well as applications where you would
not need to include the additional hardware for switch debouncing (in other words,
applications where you can tolerate switch bouncing). Note, you cannot use the clock and
clear inputs of our lab as example applications; instead you need to think of other
examples.

3. Finally, evaluate this lab exercise in terms of its value added to your understanding of
digital logic circuits. Include any suggestions you have to improve this lab for subsequent
semesters.

Turn in your written lab report (be sure to include all 3 items from the post-lab exercise
above) to your lab TA.